MADRID, April 22 (2021) –
Google Cloud has made its new cloud solutions for document processing, Document AI, which use Google’s artificial intelligence (AI) tools for actions such as document data extraction.
The new Google Cloud tools within the Document (Doc) AI platform also include DocAI for loans and DocAI for acquisitions, and they seek that companies can automatically manage their document data.
To do this, as Google informs in a statement, it uses Artificial intelligence technology, such as computer vision and natural language processing., with which companies can use predefined algorithms and train their own models, as well as the Knowledge Graph to fill in the fields of the documents and the possibility of interacting with AI to ensure accuracy.
AI document, which Google had started testing at the end of 2020 and which is now available to businesses via Google Cloud, is a document processing platform that companies can use through a unified console.
The platform is adapted with different functions according to the scope of each company, and Lending DocAI is especially dedicated to the financial services industry..
This tool processes the income data of loan recipients through attached documents, for which it uses a series of specialized machine learning models that “Automate routine document reviews”, According to Google.
The other solution is Procurement DocAI, a tool that allows companies to “accelerate the processing of documents for invoices, receipts and other valuable documents in the procurement cycle.”
The use of automated systems to capture document data in this solution allows companies to reduce your costs by up to 60 percent. Google Knowledge Graph allows you to validate and enrich the information in your documents.
To improve the accuracy of these automated systems, Google has provided them with the ‘Man-in-the-Loop AI’ function, what help aliens through human intervention and with which companies can perfect their models.
This feature works by making estimates in the form of a figure to represent the certain that the algorithm has that the data has been indexed correctly.
